My Netwok Dictionary

Definition of service quality

The term QoS (an acronym for “Quality of Service”) means the ability to provide service (such as a communication medium) which meets requirements of response time and bandwidth.

Applied to packet switching networks (based networks using routers) for QoS is the ability to be able to guarantee an acceptable level of packet loss, defined by contract, for a given use (VoIP, video conferencing, etc…).

Indeed, unlike circuit-switched networks, such as switched telephone networks, where a communication circuit is dedicated for the duration of the communication, it is impossible to predict Internet path taken by individual packets.

Thus, there is no guarantee that a communication requiring regularity of flow can take place without hindrance. Therefore there are mechanisms, called QoS mechanisms, to differentiate different network flows and reserve a portion of bandwidth for those requiring a continuous, without breaks.

Service Levels

The term service level defines the level of demand for the ability of a network to provide a service point to point or end to end with a given traffic. It generally defines three levels of QoS:

1. Best effort, providing no differentiation between flows and networks allowing no guarantee. This level of service is also sometimes called Lack of QoS.
2. Differentiated service or soft QoS, to define priority levels to different network flows without providing strict guarantees.
3. Guaranteed service or hard QoS, of reserving network resources for certain types of flows. The main mechanism used to achieve this level of service is RSVP (Resource Reservation Protocol).

Criteria for quality of service

The main criteria for judging the quality of service are as follows:

• Capacity (bandwidth): sometimes called bandwidth abuse of language, it defines the maximum amount of information (bits) per unit time.
• Jitter: it represents the fluctuation of the digital signal in time or phase.
• Latency, or delay the response time: it characterizes the delay between transmission and reception of a packet.
• Packet loss: it corresponds to the non-issuance of a packet of data, mostly due to network congestion.
• Desequencing: this is a modification of the order of arrival of packets.